Tool to Reduce Student Attention Difficulty During Online Classes Using Computational Vision

Barrial H.; Antonio O.; Xiomara F.A.D.; Alarcon J.S.D. · 35th International Conference on Computer Theory and Applications, ICCTA 2025 - Conference Proceedings (2025) · DOI: 10.1109/iccta68914.2025.11519909

Resumen

Virtual education, widely adopted following the COVID-19 pandemic, has generated significant difficulties in student concentration. Various studies reveal that a high proportion of students lose attention in the first few minutes, affecting their performance and academic engagement. Factors such as lack of direct supervision, lack of visual interaction, and the multiple distractions of the home environment affect their academic performance. Given this, this work proposes a computer vision-based tool to monitor attention in online classes and reduce attention difficulties without requiring specialized hardware. The proposal was developed in four phases: 1, selection of technological tools, prioritizing low resource consumption and local applicability; 2, visual and architectural design of the application; 3, construction of the system using Python, OpenCV, and MediaPipe; and 4, functional validation with students in real environments and with a specialist. During the experimental phase, a group of university students who take virtual classes used the tool while in class. Two tests were administered: one without alerts and another with the inattention notification feature. As a result, an average reduction of 42.3% in the dropout rate was observed, confirming the effectiveness of reducing the online student dropout rate.

Instituciones: Universidad Peruana de Ciencias Aplicadas